The ideal optically based sensor combines a high selectivity towards the species of interest, a low Limit-of-Detection (LOD), and a real time sensing capability. High selectivity can often be obtained by utilizing a narrow-band (high resolution) light source or wavelength selection detection system. Spectroscopically speaking, an optical sensor can operate in either an absorption or emission mode. Certainly, emission based sensors can produce lower LOD in many cases. On the other hand, quantifying the spectral intensities from an emission based sensor to extract information regarding species concentrations is challenging due to inherent dynamical effects (quenching, predissociation, unknown quantum yields, etc.).
Direct absorption spectroscopy methods have many experimental advantages including selectivity and ease with which the absorbance measurements can be used to quantify species concentrations. While direct absorption measurements, at least the way in which the standard infrared absorption experiments are performed, do not possess the same level of detection sensitivity as fluorescence spectroscopic methods, there are specialized measures that can be incorporated into the experiment to overcome this limitation. For example, assuming Beer's Law can be applied to the absorption measurement and a previously optimized set of experimental conditions, it should be clear that improvements in the signal to noise ratio for the measurement can be realized by increasing the absorption path length. One strategy therefore, is to incorporate a multi-pass absorption sample cell into the experiment to effectively increase the path length through the sample. Indeed, by interfacing a 32 m White cell with a standard FTIR instrument, Robitaille and coworkers have demonstrated the ability to distinguish, identify and quantify 2,4-DNT, 2,6-DNT, and TNT vapor from heated soil samples with a ppm detection sensitivity. Clapper, M., J. Demirgian, and G. Robitaille, A Quantitative Method using FTIR to Detect Explosives and Selected Semivolatiles in Soil Samples, Spectroscopy, 10(7), 44-49 (1995).
Parenthetically, the problem of soil contamination at DOD and DOE facilities apparently represents a significant environmental problem. There are a number of federally funded studies focusing on the development of down-the-hole sensors for a variety of contaminants including explosives. As another example, spectroscopists from Aerodyne Research, utilizing a tunable Pb-salt diode laser coupled to an astigmatic Herriott cell reported both laboratory mechanistic as well as in situ field studies demonstrating a sensitive, specific, real time sensing capability for TNT in soils. Wormhoudt, J., J. H. Shorter, J. B. McManus, P. L. Kebabian, M. S. Zahniser, W. M. Davis, E. R. Cespedes, and C. E. Kolb, Tunable infrared laser detection of pyrolysis products of explosives in soils, Applied Optics, 35(21), 3992-3997 (1996). The TNT soil measurements reported in the literature utilized a thermal desorption system to entrain the soil contaminants into the gas phase.
Optical sensors designed to detect vapor emissions are not limited in scope or application to explosive or energetic materials. Many illicit drugs such as heroine or cocaine are often in chloride form. As a result, chlorine-containing compounds will often be detected in the vapor emissions from these compounds. Optically based vapor sensors also have potential applications as a medical diagnostic. There are over 300 volatile organic compounds (VOC's) in exhaled breath of humans. Gordon, S. M., J. P. Szidon, B. K. Krotoszynski, R. D. Gibbons, and H. J. O'Neill, Volatile Organic Compounds in Exhaled Air from Patients with Lung Cancer, Clin. Chem., 31(8), 1278-1282 (1985). The relative concentrations of VOC's have for centuries been used to assist in diagnosis. Patients suffering from diabetes tend to have elevated levels of ketones, principally acetone, in their breath for example, and hence often smell like rotten apples. More recently, patients suffering from breast cancer have been shown to have elevated levels of formaldehyde in their exhaled breath. O'Neill, H. J., S. M. Gordon, M. H. O'Neill, R. D. Gibbons, and J. P. Szidon, A Computerized Classification Technique for Screening for the Presence of Breath Biomarkers in Lung Cancer, Clin. Chem., 34(8), 1613-1618 (1988). The challenge for optical sensors in these cases is the ability to differentiate between normal and elevated levels of VOC's in a patient.
For many applications involving energetic materials, illicit substances, or medical diagnostics however, gaining an order of magnitude or two by increasing the absorption path length to 100 m or so may still not be sufficient. Consider that high quality military explosives need not be present in large quantities to cause significant damage particularly if combined with an incendiary compound. Persons with malicious intent can further exacerbate the vapor detection problem by encasing explosives in containers specifically designed to minimize vapor emissions. For medical diagnostic applications, the most useful sensor would be one capable of detecting elevated VOC levels at a pre-symptomatic, i.e., low concentration, stage.
In the late 1980's another direct absorption method was serendipitously discovered that allows absorption path lengths of 10 kilometers to be realized. O'Keefe, A. and D. A. G. Deacon, Cavity ringdown optical spectrometer for absorption measurements using pulsed laser sources, Review of Scientific Instruments, 59(12), 2544-2551 (1988). Called cavity ringdown laser absorption spectroscopy or “CRD”, by its originators, it involves measuring changes in the characteristic ringdown time of a high Q optical cavity due to the presence of an absorbing sample. The ringdown cell is actually a type of lossmeter that was used initially to determine the reflectivity of high reflectance mirrors (R>99.9%). Over the past decade, cavity ringdown has been exploited by a number of research groups for a variety of applications. See for example, Busch, K. W. and M. A. Busch, Editors, Cavity-Ringdown Spectroscopy: An ultratrace Absorption Measurement Technique, ACS Symposium Series 720, American Chemical Society, Washington, D.C. 1999 and references therein. There have even been some preliminary studies to examine the potential of cavity ringdown for trace detection of explosive materials. Steinfeld, J. I., R. W. Field, M. Gardner, M. Canagaranta, S. Yang, A. Gonzalez-Casielles, S. Witonsky, P. Bhatia, B. Gibbs, B. Wilkie, S. L. Coy, and A. Kachanov, New Spectroscopic Methods for Environmental Measurement and Monitoring, SPIE, 3853, 28-33 (1999); Todd, M. W., R. A. Provencal, T. G. Owano, B. A. Paldus, A. Kachanov, K. L. Vodopyanov, M. Hunter, S. L. Coy, J. I. Steinfeld, and J. T. Arnold, Application of mid-infrared cavity-ringdown spectroscopy to trace explosives vapor detection using a broadly tunable (6-8 □m) optical parameteric oscillator, Applied Physics B, 75, 367-376 (2002); and Usachev, A. D., T. S. Miller, J. P. Singh, F.-U. Yueh, P.-R. Jang, and D. L. Monts, Optical Properties of Gaseous 2,4,6-Trinitrotoluene in the Ultraviolet Region, Applied Spectroscopy, 55(2), 125-129 (2001).
In the classic cavity ringdown experiment, a pulsed laser system serves as the radiation source. O'Keefe, A. and D. A. G. Deacon, Cavity ringdown optical spectrometer for absorption measurements using pulsed laser sources, Review of Scientific Instruments, 59(12), 2544-2551 (1988); and Busch, K. W. and M. A. Busch, Editors, Cavity-Ringdown Spectroscopy: An ultratrace Absorption Measurement Technique, ACS Symposium Series 720, American Chemical Society, Washington, D.C. 1999 and references therein. Output from this pulsed laser source is injected in a cavity consisting of two highly reflective mirrors (R>99.99%). Once injected into the cavity, the light pulse can traverse the cavity thousands of times, although a small portion of the intensity of the pulse leaks out of the cavity as each mirror is encountered. A detector is situated behind the mirror opposite the radiation input to monitor cavity output and/or decay of the laser pulse. For pulsed light sources, whose coherence length is short compared to the physical size of the cavity, the decay is typically exponential and possesses a decay or ringdown time characteristic for the cavity. In the presence of absorbing species, this characteristic ringdown time changes and hence absorption spectroscopy can by performed by measuring the difference in ringdown time as a function of molecular species concentration. For cavities with highly reflective mirrors, the absorption path length can approach 10 kilometers. The ultra-trace vapor detection potential of cavity ringdown is due then to this tremendous gain in path length compared with more traditional spectroscopic methods.
Prior CRD detection methods are deficient in that each of the methods take too much time to be useful in a real world environment. In the classic CRD laser experiment, an absorption spectrum, plotted as the intensity loss of the cavity versus wavelength, is actually composed of a great many individual cavity ringdown events. Collection of the absorption spectra or scan is initiated by first tuning the laser (or some other optical source) to a starting wavelength for the scan. A ringdown event is observed, averaged, and then modeled as an exponential decay in order to extract a characteristic ringdown time for the cavity at this starting wavelength and in the presence of an absorbing sample to be analyzed. Finally, this wavelength specific ringdown time, τsample(λstart), is compared with the ringdown time for an evacuated cavity, again at a specific wavelength, λstart. This difference, β=τempty(λstart)−τsample(λstart), represents the first ordered pair in the absorption spectrum (βstart, λstart). Of course an absorption spectrum includes a large number of such pairs. To continue collecting a spectrum, the laser must then be stepped or tuned to a new wavelength and the process repeated until an entire absorption spectrum has been obtained. Depending upon the size of the wavelength region to be scanned and/or the size of the individual steps for each retuning, such an experimental scheme can become quite time intensive (to the point of becoming time prohibitive for a real time sensor).
Driven by the molecular sensing potential of CRD methods, there have been a number of attempts to circumvent the time intensive nature involved with the collection of an absorption spectrum via the CRD method (which does not include the subsequent principal component analysis step required to quantify trace amounts of species in the gas sample). In some cases, a single averaged, wavelength specific, CRD event, chosen to coincide with the linecenter for a strong absorption peak in the spectrum, is used to perform the detection and quantification analysis. See for example, Wang, C., S. T. Scherrer, and D. Hossain, Measurements of Cavity Ringdown Spectroscopy of Acetone in the Ultraviolet and Near-Infrared Spectral Regions: Potential for Development of a Breath Analyzer, Applied Spectroscopy, 58(7), 784-791 (2004). Unfortunately, this strategy can severely limit the selectivity of the CRD method and, particularly for real world samples which can contain hundreds of compounds, effectively cripples the usefulness of the CRD approach. Other attempts to circumvent the time intensive collection challenge involve the use of a broadband laser or optical source. See for example, Scherer, J. J., J. B. Paul, H. Jiao, and A. O'Keefe, Broadband ringdown spectral photography, Applied Optics, 40(36), 6725-6732 (2001); and Biennier, L., F. Salama, M. Gupta, and A. O'Keefe, Multiplex integrated cavity output spectroscopy of cold PAH cations, Chemical Physics Letters, 387, 287-294 (2004). Indeed, O'Keefe and coworkers have demonstrated that such broadband light sources can in fact generate optical spectroscopic data for molecular species present only in trace amounts and in essentially real time. Scherer, J. J., J. B. Paul, H. Jiao, and A. O'Keefe, Broadband ringdown spectral photography, Applied Optics, 40(36), 6725-6732 (2001); and Biennier, L., F. Salama, M. Gupta, and A. O'Keefe, Multiplex integrated cavity output spectroscopy of cold PAH cations, Chemical Physics Letters, 387, 287-294 (2004). The trade off of course is that their broadband approach cannot achieve the same level of spectral resolution (and hence selectivity) as a narrow band laser source. In another incarnation of the broadband CRD concept, the output of a CRD cell was sent to a monochromator equipped with a diode-array or CCD detector. See Fiedler, S. E., A. Hese, and A. A. Ruth, Incoherent broad-band cavity enhanced absorption spectroscopy, Chemical Physics Letters, 371, 284-294 (2003); and Gherman, T. and D. Romanini, Mode-locked cavity-enhanced absorption spectroscopy, Optics Express, 10(19), 1033-1041 (2002). Both groups have reported broadband CRD spectra with spectral resolution on the order of several tenths of a wavenumber resolution; certainly sufficient to rotationally resolve the molecules studied in these reports (O2— Ruth and C2H2— Romanini). Unfortunately, this level of spectral resolution is insufficient to produce rotationally resolved spectra for larger molecules and, moreover, the experimental scheme described in these reports is not readily transferable to the fingerprint region of the infrared, primarily due to performance characteristics of monochromators, spectrographs, and linear array detectors.
The above discussed technical problems can be solved by the following apparatus. In essence a series of cw diode lasers, quantum cascade lasers, or other tunable laser sources, each tunable over a discrete, yet unique, fingerprint region of the infrared, will provide a capability to not only take advantage of the inherent sensitivity of the cavity ringdown method, but will also provide a high level of selectivity by allowing numerous fingerprint regions to be examined simultaneously. A PZT actuated mirror mount on each CRD cavity in the multicolor sample cell can facilitate use of these cw light sources. To ensure reliable and robust operation of the cavity ringdown instrument when interfaced with a scalable screening portal or other sampling device, hollow glass waveguides (HGW's) and/or infrared fiber optics can be utilized to interface each laser with the cavity ringdown detection cell. One of the limitations that has always been cited when comparing infrared or near infrared cavity ringdown methods with other infrared spectroscopic methods such as FTIR, is that while FTIR is orders of magnitude less sensitive, one can acquire a spectrum of the fingerprint region in less time. The invention disclosed herein effectively represents a solution to this limitation by allowing the measurement of multiple discrete fingerprint wavelengths simultaneously.